Record video and extract frames


A potentially useful feature would be the ability to click-select frames from a video, either uploaded through the uploader or ingested from a connected device, and extract those frames for training/testing data.

In my current setup, I’ve been using OpenMV to record video from my Arduino Portenta Vision Shield, and then extracting frames using VLC (YouTube video explaining how).

Hey @avielbr, thanks for the suggestion. I think it’s indeed an interesting thought, esp. if we combine it with some of the new auto-labeling tools we’re working on - would be a quick way to build out a vision dataset. Added to the backlog of ideas!